Parsons (1987) Perceived spatial organization of

We thank Ian Howard, Lawrence Marks, and an anonymous reviewer for their ... Correspondence concerning this article should be addressed to Law- rence M.
1MB taille 3 téléchargements 376 vues
Journal of Experimental Psychology: Human Perception and Performance 1987, Vol. 13, No. 3,488-504

Copyright 1987 by the American Psychological Association, Inc. 0096-1523/87/100.75

Perceived Spatial Organization of Cutaneous Patterns on Surfaces of the Human Body in Various Positions Lawrence M. Parsons and Shinsuke Shimojo Department of Brain and Cognitive Sciences Massachusetts Institute of Technology

The perceived spatial organization of cutaneous patterns was examined in three experiments. People identified letters or numbers traced on surfaces of their body when the relative spatial orientations and positions of the body surfaces and of the stimuli were varied. Stimuli on the front or back of the head were perceived with respect to a frame of reference positioned behind those surfaces, independent of the surfaces1 position and orientation. This independence may relate to the way in which the sensory apparatus on the front of the head is used in planning action. Stimuli on other surfaces of the head and body were perceived in relation to the position and orientation of the surface with respect to the whole body or trunk (most of which was usually upright). Stimuli on all transverse/ horizontal surfaces were perceived with respect to frames of reference associated with the head/ upper chest area. These frames were also used for stimuli on frontoparallel surfaces in front of the upper body. These observations may result from the use of "central" frames of reference that are independent of the head and are associated with the upper body. Stimuli on surfaces in other positions and orientations (with two exceptions) were perceived "externally"—that is, in frames of reference directly facing the stimulated surface. The spatial information processing we found may be fairly general because several of our main findings were also observed in very young children and blind adults and in paradigms studying perception by "active touch" and the spatial organization of the motor production of patterns.

In 1977, Corcoran reported intriguing demonstrations of the perception of cutaneous patterns on surfaces of the hand, head,

with their hand in front of their upper chest and head and their

and thigh. An upright 2 traced on a backward-facing back of a

the body; other investigators obtained similar results (Allen & Rudy, 1970; Duke, 1966; Krech & Crutchfield, 1958; Natsou-

head upright and facing in the same direction as the front of

subject's hand or the back of the head was perceived as a normally oriented 2, whereas an upright 2 traced on a forwardfacing palm or on the forehead was perceived as its upright mirror image. This pattern of perceptions on forward- and back-

las, 1966; Natsoulas & Dubanoski, 1964; Pedrow & Busse, 1970;Podell, 1966). These findings are of general interest because they may reveal

ward-facing body surfaces was also observed when the hand's

how the parts, surfaces, and surrounding space of the body are

orientation was reversed. An upright 2 traced on a backward-

represented for purposes of perception, action, and cognition.

facing palm was perceived as an upright 2, whereas an upright 2 traced on the forward-facing back of the hand was perceived as its upright mirror image. However, not all forward-facing

There have been recent advances in our understanding of spatial information processing associated with the production of movement (Arbib & Amari, 1985; Pellionisz & Llinas, 1980,

body surfaces showed this tendency for stimuli to be perceived

1982; Robinson, 1982; Soechting, 1982; Soechting, Lacquaniti,

as their mirror images. An upright 2 traced on a standing indi-

& Terzuolo, 1986; Soechting & Ross, 1984), and in our under-

vidual's thigh was perceived as an upright 2. All of these obser-

standing of the tactile information transmitted to, and repre-

vations were collected from standing, blindfolded individuals,

sented in, the cortex (Dellon, 1981; Johannson & Vallbo, 1979, 1983;Johnson, 1983; Johnson & Lamb, 1981; Phillips & Johnson, 1981a, 1981b). However, there have been few studies on

This research was supported by a grant from the A. P. Sloan Foundation Program in Cognitive Science to the MIT Center for Cognitive Science and by National Research Service Award postdoctoral fellowship F32-HD06605-02 from the National Institutes of Health to the first author. We thank Ian Howard, Lawrence Marks, and an anonymous reviewer for their useful comments on earlier drafts of the article, and Jeni Yamada for producing the illustrations. Shinsuke Shimojo is now at the Smith-Kettlewell Eye Research Foundation, San Francisco, California. Correspondence concerning this article should be addressed to Lawrence M. Parsons, E10-020, Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139.

488

how the perceptual system integrates and processes information from tactile receptors.

Perception of Spatial Organization and Frames of Reference In three-dimensional space, a bidimensionally asymmetric pattern such as a letter or number is intrinsically ambiguous because it can be described within more than one frame of reference. For example, it can be described with respect to an intrinsic (or object-centered) frame of reference, as shown in Figure 1 (Hinton & Parsons, 1981; Marr & Nishihara, 1978). It can also be described with respect to an extrinsic frame of refer-

PERCEIVED SPATIAL ORGANIZATION

489

top

front

front

leading side

leading side

leading 9ide

top

front top

Figure 1. Intrinsic, object-centered frames of reference: An upright and upside-down R described in an intrinsic frame and shown from the front and back. (The far right figure is an upside-down, mirror-reversed R shown from the front. The difference between the descriptions of two upside-down letters is that the middle one is described in a left-handed frame and the right one is described in a right-handed frame.)

ence—that is, one centered at some location other than on the object, as shown in Figure 2. Researchers recognize the role of three frames of reference in the perception of shape and spatial relations (Figure 3): that of the perceiver, the perceived object, and the (local) environment containing them (Clowes, 1969; Hinton, 1981; Hinton & Parsons, 1987; Palmer, 1977; Pinker, 1985; Rock, 1973;Sedgwick, 1983; Shepard & Hurwitz, 1984). These three frames of reference are also useful in analyzing the perceived spatial organization of cutaneous patterns on surfaces of the body. However, the situation is complicated because parts and surfaces of the body can adopt many different configurations (cf. Committee for the Study of Joint Motion, American Academy of Orthopaedic Surgeons, 1965). Generally, the perceptual system must determine the mapping among the possible K orientations (with respect to the environment) of a sensory surface, L configurations of the body, M orientations of the body as a whole with respect to the environment, and jV frames of reference that may describe a stimulus. A system could be effective while using only limited regions of this KY.Ly.My. .W space of possible mappings if it followed a few principles. Findings like Corcoran's (1977) suggest that this is what people do. There are three kinds of interpretative systems, each with many possible variants. In an environment-based system, cutaneous stimuli are referred to a coordinate system organized either universally (e.g., in directions implied by gravity and a compass) or locally (implied by local landmarks). (This system is not investigated here but in Parsons & Shimojo, 1987.) In a local surface-based system, a stimulus is referred to intrinsic features of the stimulated surface (its frame of reference) without regard for the surface's orientation and position (Figure 4 A and 4B). (A local surface-based system is embodied in neurological exams, in which the patient is expected to report the identity of cutaneous stimuli from the point of view of an examiner who directly faces the stimulated surface.) In a whole bodybased system, a stimulus is described in a frame of reference assigned to the trunk or whole body. This frame of reference may be positioned, oriented, and organized in various ways (i.e., using one of the many possible coordinate systems and metrics—see, e.g., Morse & Feshbach, L953, p. 655 ff.). A surface's position and orientation with respect to the trunk or whole

body determines the perceived spatial organization of cutaneous stimuli (Figure 4C and 4D). Assuming that there is a frame for each principal surface of each body part, a local surface-based system requires about 56 local frames of reference. A whole body-based system requires only a single frame of reference. However, this latter system needs sensorimotor information (cf. Clark & Horch, 1986) about the configuration of the relevant body parts in order to "compute" the coordinates of cutaneous stimulation with respect to this frame of reference. Computing this relation directly or initially facilitates later comparisons of the spatial properties of cutaneous stimulation received at different surfaces or at the same surface but at different times (i.e., with the body, its parts, and the stimulus in different configurations). The information perceived serially from partial cutaneous stimuli could be used to build a representation of the whole object. (This is similar to integrating information gathered from successive glances at an object or scene—see, e.g., Hochberg, 1968; Turvey, 1977.) The comparison of stimulation received at different moments or at different skin surfaces is likely to be more difficult in a system that uses only local frames of reference because it requires access to representations of configurations of the body, its parts, and the stimulus at arbitrary moments in the past. If the perceptual system requires frequent comparison or integration of spatial information received at different instants or different surfaces, a whole body-based system is more efficient than a local surface-based one. The topology of surfaces containing an organism's sensory and motor processes, as well as functions of the perceptual system other than those just discussed, may also determine the optimal interpretative system.1 If peripheral components of the sensory or motor systems are concentrated on some surfaces, those surface might be "privileged," because action and perception are organized or coordinated with respect to those surfaces. Local or special frames of reference might be used for such privileged surfaces, whereas (for the reason described earlier) a whole body-based system might be used for other surfaces. 1 See, for example, Hecaen and Albert (1978, p. 292 ff.) and Ratchff (1982) for reviews of solutions to the problem of reorganizing such a system after injury to the nervous system or body.

490

LAWRENCE M. PARSONS AND SH1NSUKE SHIMOJO

Figure 2. Letter R with respect to frames of reference centered at positions other than on the R.

Experimental Paradigms and Findings The position and orientation of the reference frame used to interpret a stimulus on a surface may be revealed by examining how that interpretation is affected by changes in the disposition of a stimulus on a surface and in the disposition of a stimulated surface relative to the whole body (Figure 4). Such methods could signal use of a local surface-based or whole body-based system. The latter method has been used with limited effectiveness. Results from such studies suggest that only the head's surfaces are described in local frames of reference, whereas other surfaces are described in one or another frame dependent on the body's intrinsic frame (or on a frame that is a compromise among intrinsic frames of the head, trunk, lower limbs, and local environment). Egocentric and environmental frames have been confounded in nearly all preceding work, so it is not known in which of the two frames such stimuli are described: See Oldfield and Phillips (1983) and the General Discussion section. Previous findings, though useful, are insufficient or too ambiguous to support a full and detailed description of this system. Changes in position have often been confounded with changes in orientation, body part, or body surface; furthermore, few surfaces, orientations, and positions have been examined. The purpose of the studies we report is to provide a systematic empirical basis for understanding individuals' systems of interpretation for cutaneous stimuli. We first compare perception of cutaneous stimuli at many surfaces of the body in a natural standing position. Then we examine how the perception of cutaneous stimuli on a body surface is influenced by the orientation and position of the surface. (We do not investigate whether for a given spatial relation of the stimulated surface to the whole body or trunk, perception is affected by the configuration of the unstimulated body parts.) We also examine the perceptual system's preference for interpreting stimuli that are multiply ambiguous. A stimulus in Experiments 2 and 3 is ambiguous not only with respect to its "handedness" (whether it is a mirror reversed or normal form of the experimenter-defined

stimulus) but also with respect to the orientation of its topbottom axis. For brevity, we refer to the experimenter-defined handedness of a stimulus as external handedness; we call the oppositehanded version of a stimulus mirror reversed.

Experiment 1: Stimuli on Body Surfaces in a Natural Standing Position In Experiment 1 we surveyed cutaneous pattern perception on many body surfaces of a standing, blindfolded individual. A normal or mirror-reversed 2. k. or L was traced on the surfaces shown in Figure 5. These surfaces were meant to be representative of the whole body. We assumed, as did others, that the perception of spatial organization of cutaneous stimuli on body parts of the left half of the body was indistinguishable from that on body parts of the right half, and therefore we did not com-

Figure 3. Examples of frames of reference for the percerver, perceived object, and local environment.

PERCEIVED SPATIAL ORGANIZATION

491

pare the left and right halves.1 All but four stimulated surfaces were vertical; the others—the top of the tongue and head, and the front and back surfaces of the top of the shoulder—were nearly horizontal. On vertical surfaces, a letter or number was traced with its top oriented upward (toward the top of the body). Stimuli on the top of head and the back of the top of shoulder were traced with their tops pointing forward (facing in the same direction as the front of the body and head). Stimuli on the top of tongue and the frontal surface of the top of the shoulder were

Figure 5. The position and orientation of the body and the surfaces stimulated in Experiment 1.

traced with their tops pointing backward (facing in the opposite direction that the front of the body and head). The subjects were instructed to report their very first perception of the identity and handedness of the stimulus.

Method

Figure 4. Panels A and B: Example of a local surface-based frame of reference that could be used to describe a stimulus (a 2 with its top toward the wrist) traced on the back of the hand in two different positions and orientations. (In both cases, the stimulus is described as an upright normal version of a 2.) Panels C and D: A whole body-based frame of reference used to describe a 2 (with its top toward the wrist) traced on the back of the hand in two different positions/orientations in front of the body. (The frame is positioned with its top-bottom axis along the spine. In Case C, the stimulus is described as an upright mirror-reversed 2; in Case D, it is described as an upside-down normal 2.)

Subjects. Eighteen male undergraduates at the Massachusetts Institute of Technology, who had not been in any related experiments, received $5.25 per hour for participating in this study. Stimuli. The characters 2, h, L, or their mirror reversals were traced with a soft, plastic-tipped stylus either on the skin or on the overlying clothes (Figure 5). Stimuli were as close as possible to 7 x 7 cm in size, and on every trial they were traced with identical strokes, moving from the top of the character downward. The duration and force of the strokes were uniform across characters, surfaces, and trials. The same experimenter traced the characters for all subjects in all the experiments reported here. Design. Two sets of 66 trials were used, with half of the subjects randomly assigned to perform each set. In each set, half of the characters were mirror reversed and half were normal, and one or the other version of a 2, h, or L was traced on each of 22 body surfaces. A surface never received either all mirror reversed or all normal characters, and it was never stimulated on consecutive trials. The two sets of trials exhausted (without duplication) all passible combinations of body surface, stimulus identity, and stimulus handedness. The order of trials was random and unique for each subject. Procedure. Subjects first visually inspected ink-drawn versions of six stimuli. Then they stood with eyes covered and performed six practice 2 However, for evidence that some aspects of the information processing in tactile-spatial tasks differ with respect to the left and right sides of the body, see, for example, DeRenzi, 1982; Hermelin and O'Connor, 1971; Rudel, Denckla, and Hirsh, 1977; Rudel, Denckla, and Spalten, 1974; Smith, Chu, and Edmonston, 1977.

492

LAWRENCE M. PARSONS AND SHINSUKE SHIMOJO

trials with each stimulus drawn on the back of their head and then on their forehead. They were told to report their very first impression of the identity and handedness of each character. It was emphasized that there was no single correct answer, that the experimenter's point of view was not pertinent in any way, and that they need not be consistent across trials but should simply report their spontaneous interpretation of each stimulus as quickly as possible. Then two sets of 33 test trials were performed, with an intervening 5-min break. A trial began with the experimenter's indicating where a stimulus would be traced, and the trial ended with the subject's making a verbal report (e.g., "L, normal"). Subjects could request repeated presentations of a stimulus, but this occurred on less than 15% of trials.

Results There was no significant difference between the responses of the two groups of subjects, as determined by a chi-square test of the responses for each surface at each orientation and position; therefore, groups were combined in all further analyses. Subjects reported perceptions that did not always match the external handedness of the stimulus but that nearly always matched its identity.3 On all surfaces except the forehead, tongue, and top of the foot, subjects reported perceiving stimuli as matching the external handedness. (For each surface at each orientation and position, the chi-square test against chance performance ranged from 24.0 to 4.16, with df= \,N= 54, p < .05.) Stimuli on the top of the foot were also reported to have the external handedness more frequently than the reverse, but only marginally so. Table 1 summarizes the responses of 16 subjects and shows the responses of 2 other subjects, which were markedly different from the majority and are discussed separately. Forehead and tongue. Subjects consistently perceived mirror reversals of stimuli traced on the forehead, whereas they perceived stimuli traced on the back and top of the head as having external handedness. Subjects were less consistent in their responses to stimuli on the tongue: They mostly perceived stimuli on the underside of the tongue as mirror reversals, but this ten-

Table 1 Experiment 1: Handedness of Perceived Character on Various Surfaces of Body in a Natural Standing Position (N= 18) Percentage of responses matching experimenter-defined handedness Surface Head Forehead Back Top Trunk Front Right side Back Right hand Palm Back Right thigh Front Back Inside Outside Right upper arm Front Back Outside Inside Top of right shoulder Backward Forward Right foot Top

Sole Tongue Underside Top

16 subjects

2 deviant subjects

20.8" 100.0*** 95.8***

50.0 83.3 66.6

70.8 91.6*** 95.8*'*

16.6 16.6 83.3

75.0* 75.0*

66.6 16.6

85.4*** 83.3** 89.5*** 93.7***

50.0 66.6 100.0 33.3

83.3** 91.6*** 81.2** 77.0**

33.3 83.3 16.6 66.6

93.7*** 79.1**

83.3 33.3

64.5 72.9*

33.3 66.6

37.5 58.3

0 0

* Reliably different from that expected by chance (as tested by a chi square,^ l,N= 54p